The techniques for fabricating integrated circuits have been developed into the sub-micron level. Therefore, due to the trend of progressively minimizing the dimensions of the integrated circuits in the art, the silicides having therewith a relatively low resistivity are widely utilized in local interconnection, gate, and source and drain contact areas for reducing the series resistance and contact resistance. TiSi.sub.2, CoSi.sub.2, and PtSi are the most emphasized silicides attracting artisans in the art nowadays. They have their individual advantages, and shortages as well, on the applications to the integrated circuits. Both of the CoSi.sub.2 and PtSi have advantages of low resistivity, low silicidation temperature, chemical stability, inactivity to doped atom such as As or B, and characteristics of self-alignment on processing. CoSi.sub.2 is capable of forming an epitaxy on a (100) crystalline silicon, as disclosed in Ref. 1. Nevertheless, some shortages of these silicides still exist.
First of all, Al is a hot and outstanding material as a conducting medium to be deposited on the silicide, e.g. CoSi.sub.2 or PtSi, for a connection to an outer conducting wire. However, the reaction temperature of either CoSi.sub.2 or PtSi with Al is relatively low. The temperature required for reacting CoSi.sub.2 with Al is about 400.degree. C. and that for PtSi is only about 250.degree. C. The low reaction temperature will result in adverse influences to the subsequent heating process after the step of Al deposition since the silicide will react with Al if the temperature of the post-Al process used in the heat treatment is higher than the temperature at which the reaction would occur. For eliminating the adverse influences, a feasible way is to deposit a barrier layer, e.g. TiW layer, after the formation of the silicide, and then deposit the Al layer on the barrier layer. The TiW barrier layer can thereby inhibit the reaction between the silicide and Al.
Secondly, as cobalt deposited on Si is liable to be oxidized during a silicidation annealing, the silicidation process of cobalt must be executed in a high temperature vacuum fumace or by a rapid thermal annealing (RTA), as disclosed in Refs. 2-5. However, this requires equipments of high costs and is not compatible with the processing equipments presently employed in the production line. Another alternative is to coat an inert protective layer on the Co layer before the silicidation process, as disclosed in Ref. 6. However, it takes much more steps and complicates the entire processing.
Thirdly, the large consumption of Si during the silicidation of cobalt is harmful to a shallow silicided junction. For the silicidation of cobalt to form CoSi.sub.2, Si consumption per unit .ANG. of cobalt is about 3.64.ANG., as disclosed in Ref. 7. If the thickness of the Co layer is too large, then the shallow junction will be totally destroyed. Conventionally, a thin Co layer is deposited on the Si layer to prevent excessive consumption of Si and reduce the thickness of the silicide as well so as to preserve the shallow junction. Nevertheless, if a thin metal film is deposited on the Si layer, pin holes will easily occur thereon, which is notoriously adverse to the properties of the product.
For solving the aforementioned problems for the applications of silicides, a two-metal alloy scheme was disclosed, as disclosed in Ref. 8. This method utilizes two metals having different silicidation temperatures, which are usually one noble metal such as Pt or Pd and one refractory metal such as W, to be simultaneously deposited on the Si substrate. Due to the difference of the silicidation temperatures of the two metals, when an appropriate temperature lying between the silicidation temperatures of the two metals is applied to the annealing process, the noble metal having a lower silicidation temperature will be silicidized to form a silicide layer directly in contact with the Si substrate while the refractory metal having a higher silicidation temperature will not be silicidized and is isolated from the Si substrate by the silicide layer. The resulting structure would be W/PtSi/Si or W/PdSi/Si. Therefore, the W layer isolated from the Si substrate can act as a diffusion barrier layer between the silicide and the Al layers.
Although the two-metal alloy scheme would possibly solve the above-mentioned problems faced by the applications of the silicide, the effect and the performance are not as good as what are expected. First, the diffusion barrier property of the W layer, which acts as a diffusion barrier layer, is not ideal. Second, a heat treatment therefore must be executed in a vacuum furnace or in a helium furnace, as disclosed in Refs. 9 and 10, which is inconvenient and not ideal. Recently, there is one process for simultaneously forming a TiN barrier layer and a CoSi.sub.2 layer. However, the heat treatment is executed by a rapid thermal annealing process, which is not matchable with the nowaday fabrication equipment such as a nitrogen furnace.
The references cited by the Inventor are summarized as follows and hereinbefore called Refs. 1-10 respectively.
1. M. L. A. Dass, D. B. Fraser, and C. S. Wei, Appl. Phys. Lett., 58, 1308 (1991).
2. A. E. Morgan, E. K. Broadbent, M. Delftno, B. Coulman, and D. K. Sadana, J. Electrochem. Soc., 134, 925 (1987).
3. L. Van den hove, R. Wolters, K. Maex, R. De Keersmaecker, and G. Declerck, J. Vac. Sci. Technol., B4, 1358 (1986).
4. M. Tabasky, E. S. Bulat, B. M. Ditchek, M. A. Sullivan, and S.C. Shatas, IEEE Trans. Electron Device, ED-34, 548 (1987).
5. F. M. Yang and M. C. Chen, Thin Solid Films, 207, 75 (1992).
6. F. M. Yang and M. C. Chen, J. Vac. Sci. Technol., B9, 1497 (1991).
7. S. P. Murarka, Silicides for VLSI Applications (Academic, New York, 1983), p. 130.
8. S. P. Murarka, Silicides for VLSI Applications (Academic, New York, 1983), p. 169.
9. M. Eizenberg and K. N. Tu, J. Appl. Phys., 53, 1577 (1982).
10. G. Ottaviani, K. N. Tu, J. W. Mayer, and B. Y. Tsaur, Appl. Phys. Lett., 36, 331 (1980).